Читать книгу Randomised Clinical Trials - David Machin - Страница 113

The number of subjects required will depend on the hypotheses to test, the trial design, the type of endpoint variables and whether or not characteristics of the patients themselves need to be taken into account. An important consideration when deciding the eventual trial size is whether such a recruitment target can be achieved in a reasonable time frame. In many instances, developing teams are overly optimistic in this regard.

Although some details of the trial size process are somewhat routine, in that conventional wisdom dictates that the (two‐sided) test size is set minimally at 5% and the power minimally at 90%, other details such as the anticipated effect sizes should be the subject of long discussion amongst the protocol development team. Neither should they accept the conventional wisdom just indicated without debate. Test size is discussed in Chapters 8 and 9 and power in Chapter 9.